Optimal entry deterrence under uncertainty